Project Summary Mycobacterium tuberculosis is the cause of human tuberculosis, which results in pulmonary and disseminated infection. Each year TB infection results in 1.3 million deaths, while it is estimated that 1/3 of the world?s population is latently infected. When MTB infects the lung, it is typically sequestered in complex aggregations of immune cells and fibroblasts known as granulomas. However, immune responses within individual granulomas often lead to divergent outcomes. Within a given individual, many lesions can simultaneously exist in multiple states ? some lesions are able to control bacterial replication while others support persistent bacterial growth eventually leading to the spread of disease. It is likely that these differences in bacterial control across granulomas arises from combined differences in cell-type composition, cell-intrinsic activation states, and cell- cell interactions within each granuloma. Until recently, disentangling this level of complexity seemed insurmountable. However, we have recently developed and applied innovative technology for single-cell mRNA sequencing to profile restrictive and permissive non-human primate (NHP) granulomas of known bacterial burden at single-cell resolution. High-dimensional single-cell transcriptional profiling allows unparalleled resolution of multi-cellular communities. To date, we have recovered transcriptional transcriptomes of over 200,000 single cells from 40 MTB granulomas from 6 animals. Crucially, since the granulomas were harvested at 10 weeks post-infection when bacterial burden had just begun to decline in restrictive lesions, we believe that the differences in immune ecosystems between these lesions will be causally related to bacterial control. Here, I propose to combine computational and experimental approaches to test the hypothesis that differences in cell type abundance, phenotypic identity and cell-cell signaling networks correlate with bacterial control in MTB granulomas. Initially, I will construct a map of cell type diversity across granulomas and examine whether differences in cell-type composition predicts granuloma-level bacterial control. I will then explore how phenotypic diversity among macrophages within MTB granulomas influences bacterial control. Finally, I will use novel computational analyses to examine patterns of cell-cell interactions across granulomas that I will experimentally validate using in situ imaging and in vitro perturbation. If successful, the proposed analysis and experiments will provide an unprecedented understanding of the immune correlates of MTB control at the level of individual disease lesions. I envision that this study has potential to identify previously unappreciated cell types and diversity across MTB granulomas and nominate novel strategies for host-directed therapy and prophylaxis in MTB infection.